This is the proudest release of my career :) At Liquid AI, we're launching three LLMs (1B, 3B, 40B MoE) with SOTA performance, based on a custom architecture. Minimal memory footprint & efficient inference bring long context tasks to edge devices for the first time!

It is interesting. Although I've created similar notebooks for myself as experimental trial.

BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Physics to John J. Hopfield and Geoffrey E. Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”

Karapathify 😆 Cool stuff. Going to check it out.

Spent my birthday evening at a secretive data Meetup—no photos allowed, so no pics here! Just know it was special and interesting. It was fun and inspiring.

I don’t know if we live in a Matrix, but I know for sure that robots will spend most of their lives in simulation. Let machines train machines. I’m excited to introduce DexMimicGen, a massive-scale synthetic data generator that enables a humanoid robot to learn complex skills

*Do LLMs learn to reason, or are they just memorizing?*🤔 We investigate LLM memorization in logical reasoning with a local inconsistency-based memorization score and a dynamically generated Knights & Knaves (K&K) puzzle benchmark. 🌐: memkklogic.github.io (1/n)

Claude 3.5 Haiku is now available on our API, Amazon Bedrock, and Google Cloud's Vertex AI. Haiku is fast and particularly strong at coding. It outperforms state-of-the-art models—including GPT-4o—on SWE-bench Verified, which measures how models solve real software issues.

BREAKING NEWS: Secret acquirer of $15+ million domain chat .com revealed and it's exactly who you'd think. For those of you that have been following me for a while, you may recall that I announced earlier this year that I had acquired the domain chat .com for an "8 figure sum"

Me: based on what you know about me. draw a picture of what you think my current life looks like. ChatGPT:

Moravec's paradox in LLM evals I was reacting to this new benchmark of frontier math where LLMs only solve 2%. It was introduced because LLMs are increasingly crushing existing math benchmarks. The interesting issue is that even though by many accounts (/evals), LLMs are inching

Watch this talk to immerse yourself into Patrick's insights on AI's impact on software. Give yourself time to absorb it—sleep on it, even. He presents many thought-provoking points, each worthy of hours, if not days, of reflection and exploration.

It's common sense and applies way beyond software.

It is simple yet profound.

I should keep repeating this for myself.

Shoutout to Radek Osmulski for the incredible work on the Meta Learning book!  Your insights go beyond ML—packed with tactical tips that make learning effective & enjoyable. You walk the talk, sharing your experience and giving back to the community.  Highly recommended!